Crack recognition automation in concrete bridges using Deep Convolutional Neural Networks

Using Unmanned Aerial Systems (UASs) for bridge visual inspection automation necessitates the implementation of Deep Convolutional Neural Networks (DCNNs) to process efficiently large amount data collected by UASs sensors. However, these networks require massive training datasets defects recognition and detection tasks. In an effort expand existing concrete datasets, particularly cracks in bridges, this paper proposes a public benchmark annotated image dataset containing over 6900 images cracked non bridges culverts. The presented includes some challenging surface conditions covers with different sizes patterns. authors analyzed proposed using three state art DCNNs Transfer Learning mode. models were used classify best testing accuracy obtained reached 95.89%. experimental results showcase potential use train deep crack bridges. is publicly available at https://github.com/MCBDD-ZRE/Concrete-Bridge-Crack-Dataset- academic purposes.